LaneNet: Real-Time Lane Detection Networks for Autonomous Driving

• DOI: 10.48550/arxiv.1807.01726
• Publication Date: 2018-01-01
• Authors: Ze Wang; Weiqiang Ren; Qiang Qiu
• Keywords: Line (geometry); Line segment

Lane detection is to detect lanes on the road and provide accurate location shape of each lane. It severs as one key techniques enable modern assisted autonomous driving systems. However, several unique properties challenge methods. The lack distinctive features makes lane algorithms tend be confused by other objects with similar local appearance. Moreover, inconsistent number a well diverse line patterns, e.g. solid, broken, single, double, merging, splitting lines further hamper performance. In this paper, we propose deep neural network based method, named LaneNet, break down into two stages: edge proposal localization. Stage uses for pixel-wise classification, localization in stage then detects proposals. Please note that goal our LaneNet built only, which introduces more difficulties suppressing false detections marks like arrows characters. Despite all difficulties, shown robust both highway urban scenarios method without relying any assumptions or patterns. high running speed low computational cost endow capability being deployed vehicle-based Experiments validate consistently delivers outstanding performances real world traffic scenarios.